Available online at www.sciencedirect.com

Psychiatry Research 157 (2008) 159–168 www.elsevier.com/locate/psychres

Personality dimensions in obsessive–compulsive disorder: Relation to clinical variables☆ ⁎ Pino Alonso a, , José M. Menchón a, Susana Jiménez a, Jacint Segalàs a, David Mataix-Cols b, Nuria Jaurrieta a, Javier Labad a, Julio Vallejo a, Narcís Cardoner a, Jesús Pujol c a Obsessive–Compulsive Disorder Clinical and Research Unit, Department of Psychiatry, Hospital Universitario de Bellvitge, Barcelona, Spain b Division of Psychological Medicine, GKT School of Medicine and Institute of Psychiatry, London, UK c Centre d'Alta Tecnologia de Barcelona. Parc de Recerca Biomèdica de Barcelona, Spain Received 17 February 2005; received in revised form 9 January 2006; accepted 14 June 2006

Abstract

Research on the relationship between personality factors and obsessive–compulsive disorder (OCD) has proved difficult to interpret due to conceptual problems including a lack of consensus on the model of personality employed as a framework as well as a failure to consider the clinical heterogeneity of the disorder. The aim of this study was to examine the dimensional personality profile associated with OCD and to determine whether any relationship exists between personality factors and clinical variables in a sample of 60 OCD outpatients who were administered Cloninger's and Character Inventory (TCI). The Yale–Brown Obsessive–Compulsive Scale (Y-BOCS), the Hamilton Rating Scale (HDRS) and the Y-BOCS symptom checklist were used to assess the severity of obsessive–compulsive and depressive symptoms and the presence of the main OCD symptom dimensions. OCD patients showed significantly higher scores in and lower scores in novelty-seeking, self- directedness and cooperativeness than healthy subjects. These results remained unchanged when only pure OCD patients without comorbid psychiatric conditions were considered. Comorbid depressive symptoms and hoarding obsessions and compulsions were significantly associated with high harm avoidance scores. These results support the existence of a dimensional personality profile associated with OCD and characterized by high harm avoidance and low novelty-seeking, self-directedness and cooperativeness scores, but also emphasize the importance of considering the influence of comorbid clinical conditions or symptom subtypes in addressing the role of personality factors in OCD. © 2006 Elsevier Ireland Ltd. All rights reserved.

Keywords: Temperament and Character Inventory; Depression; Obsessive–compulsive symptom dimensions; Hoarding

1. Introduction ☆ Presented in part at the 32nd EABCT Congress; September 20, 2002; Maastricht, The Netherlands. Although personality factors such as indecisiveness ⁎ Corresponding author. Department of Psychiatry, Hospital Uni- or orderliness were postulated to play a fundamental role versitario de Bellvitge, c/ Feixa Llarga s/n, 08907 Hospitalet de – Llobregat, Barcelona, Spain. Tel.: +34 93 2607659; fax: +34 93 in the development of obsessive compulsive disorder 2607658. (OCD) by Janet (1903) and Freud (1908) one century E-mail address: [email protected] (P. Alonso). ago, a unanimous view on the relationship between

0165-1781/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2006.06.003 160 P. Alonso et al. / Psychiatry Research 157 (2008) 159–168 personality and OCD remains elusive. Results from OCD or previous duration of the disorder has not been studies on categorically defined personality disorders addressed to date. are quite heterogeneous, with comorbid Axis II diag- Little attention has been paid to differences in noses ranging from 33% to 83% and involving such personality features among OCD subtypes. Horesh et al. different personality disorders as dependent, histrionic (1997) reported that patients with contamination obses- and schizotypal (see Summerfeldt et al., 1998, for a sions and cleaning compulsions were much more likely to review). A dimensional perspective may be a more have personality disorders than those with predominantly useful approach in describing personality functioning checking rituals, although these results were not replicat- and may better reveal meaningful relationships between ed by Matsunaga et al. (2001). Symmetry and hoarding personality and OCD. A psychobiological model of obsessions and compulsions have been described as being personality, which includes four temperament dimen- associated with obsessive personality traits (Baer, 1994), sions, termed novelty-seeking (NS), harm avoidance and hoarding significantly predicted the presence of (HA), reward dependence (RD) and persistence (P), and cluster C personality disorders (Mataix-Cols et al., 2000). three character dimensions labelled self-directedness In a recent separate study of 56 OCD patients, our group (SD), cooperativeness (C) and self-transcendence (ST), detected that hoarding symptoms in OCD were associated has been proposed by Cloninger and associates with higher scores on Gray's Sensitivity to Punishment (Cloninger, 1987; Cloninger et al., 1993). Temperament scale and lower scores on Eysenck's Psychoticism scale dimensions are defined by individual differences in (Fullana et al., 2004). To our knowledge, the relationship associative in response to novelty, danger and between Cloninger's temperament and character features reward. They are independently heritable, manifest early and OCD symptom dimensions has not been previously in life and are hypothesized to be related to specific examined. systems. Character dimensions are described in terms of response biases related to different 1.1. Aims of the study concepts of the self and mature in adulthood. Few studies concerning Cloninger's biogenetic theory The purpose of this study was (1) to evaluate the of personality have been conducted hitherto in OCD patterns of biogenetic temperament and character in 60 patients. All of them have reported higher scores on the OCD patients compared with a group of 60 age-, sex- and harm avoidance dimension in obsessive subjects com- education-matched healthy controls, (2) to determine pared with healthy volunteers (Pfohl et al., 1990; Richter whether personality dimensions assessed by the Tem- et al., 1996; Bejerot et al., 1998; Kusunoki et al., 2000; perament and Character Inventory (TCI) were related to Lyoo et al., 2001). Other reported temperament deviations clinical variables such as severity of OCD and depressive include low novelty-seeking (Kusunoki et al., 2000; Lyoo symptoms, age at onset or previous duration of the et al., 2001) and high reward dependence (Pfohl et al., disorder or presence/absence of comorbid conditions, 1990). Regarding patterns of character, low scores on the and (3) whether any relationship exists between patterns self-directedness (Bejerot et al., 1998; Kusunoki et al., of temperament and character and previously identified 2000; Lyoo et al., 2001) and cooperativeness (Bejerot et OCD symptom dimensions. al., 1998; Kusunoki et al., 2000) dimensions have been observed. Nevertheless, most of these studies fail to 2. Method consider the possible influence of the sample's clinical status in the assessment of personality. Some personality 2.1. Subjects domains, specifically harm avoidance, have been de- scribed as being related to depressive mood states in Subjects comprised 60 outpatients with OCD (37 male normal volunteers (Svrakic et al., 1992), patients with and 23 female) and 60 age-, sex-, and education-matched first-episode psychosis (Strakowski et al., 1992)and healthy comparison subjects. OCD subjects were subjects with major depression (Joffe et al., 1993). Only recruited from the OCD Clinic of Bellvitge University the study conducted by Lyoo et al. (2001) analysed the Hospital (Barcelona, Spain) between 1999 and 2001. Of influence of OCD and depression severity on personality these patients, 38 had participated in a long-term follow- patterns, while all other reports were obtained from both up study of OCD outcome and had undergone standard- symptomatic (Pfohl et al., 1990; Richter et al., 1996)and ized treatment in our unit for a minimum period of 1 year recovered (Bejerot et al., 1998; Kusunoki et al., 2000) previous to personality assessment, while 22 were new patients. Finally, the relationship of personality dimen- consecutive outpatients referred for OCD treatment to our sions to other clinical variables such as age at onset of Hospital during this period. All patients met DSM-IV P. Alonso et al. / Psychiatry Research 157 (2008) 159–168 161 criteria for OCD (American Psychiatric Association, Brown Obsessive–Compulsive Scale (Y-BOCS) (Good- 1994) and had had symptoms of OCD for at least 1 year. man et al., 1989) which establishes the following Diagnosis was independently assigned by two psychia- severity levels: subclinical (score of 0–7), mild (8–15), trists (P.A. and J.M.M.) with extensive clinical moderate (16–23), severe (24–31) and extreme (32–40). experience in OCD, who separately interviewed the A clinician-administered version of the 21-item Hamil- patients using the Structured Clinical Interview for DSM- ton Depression Rating Scale (HDRS) (Hamilton, 1960) IVAxis I Disorders–Clinician Version (SCID–CV) (First was used to assess the severity of depressive symptoms et al., 1997). Patients were eligible when both research (scores from 0 to 63). examiners agreed on all criteria. Exclusion criteria were Recent theories on the multidimensional and hetero- past or present history of psychoactive substance abuse, geneous nature of OCD (see Mataix-Cols et al., 2005,for age under 18 or over 65 years, mental retardation and a review), which attempt to search for OCD symptom severe organic or neurological pathology, except tic dis- dimensions rather than categorizing patients into mutually order. Comorbidity with other DSM-IV Axis I disorders exclusive subgroups, were adopted in our study. Despite was not considered an exclusion criterion provided OCD some differences, factor-analytical studies have been was the main diagnosis and the primary reason for seeking fairly consistent in reducing obsessive symptoms into medical assistance. During the selection period, 79 three to five clinically meaningful dimensions (Baer, outpatients were assessed by the examiners and fulfilled 1994; Leckman et al., 1997; Mataix-Cols et al., 1999; DSM-IV criteria for OCD. Of these patients, eight were Summerfeldt et al., 1999), which, at least in adult patients, ruled out in accordance with the exclusion criteria and 11 tend to remain stable over time (Mataix-Cols et al., 2002). refused to take part in the study. In this study, a clinician-administered version of the Y- Patients from the long-term follow-up study (n=38) BOCS Symptom Checklist (Goodman et al., 1989), had undergone standardized treatment for OCD in our which comprises a list of more than 50 examples of unit for 27.5±10.2 months (range=12–45), including obsessions and compulsions, was employed to ascertain exhaustive pharmacological and cognitive–behavioural scores on five previously identified symptom dimensions therapy – details of treatment strategies in this group of (Mataix-Cols et al., 1999)designatedas“Symmetry/ patients are reported in Alonso et al. (2001). At the time Ordering”, “Hoarding”, “Contamination/Cleaning”, “Ag- of the personality assessment, they were all receiving gression/Checking”,and“Sexual/Religious obsessions”. pharmacological treatment. The remaining 22 newly If a patient identified at least one of the specific symptoms referred patients were all medication free for at least under one of these dimensions as a principal or major 2 weeks at the time of the assessment. problem, that dimension was assigned a score of 2. If a Healthy comparison subjects, recruited from residents patient endorsed at least one of the specific symptoms but of the local community, were asked to participate in a did not consider it to be a major problem, that dimension study on psychological health with no payment offered. was assigned a score of 1. Finally, a score of 0 was They had no past or current history of psychiatric or assigned if a patient did not endorse any of the symptoms neurological diagnoses as determined in a brief interview under that dimension. based on the Structured Clinical Interview for DSM-III-R: Non-Patient Version (SCID-NP) (Spitzer et al., 1989)and 2.3. Personality assessment the guidelines established by Shtasel et al. (1991) to exclude psychiatric disorders. The Temperament and Character Inventory (TCI) Written informed consent was obtained from each version 9 (Cloninger et al., 1994) was used to measure subject after complete description of the study, which biogenetic temperament and acquired character. The TCI was approved by the Institutional Review Board. is a self-administered paper-and-pencil, true–false ques- tionnaire of 240 items constructed by Cloninger et al. to 2.2. Clinical assessment assess four basic dimensions of temperament, namely novelty-seeking, harm avoidance, reward dependence Information was obtained on sociodemographic – and persistence, and three primary dimensions of age, sex, and years of education – and clinical variables character, namely self-directedness, cooperativeness and – age at onset of OCD (defined as age when symptoms self-transcendence. Each of the seven dimensions, except became a significant source of distress and interfered persistence, has three to five lower-order subscales; with the patient's social functioning) – and previous because they are not usually considered in the literature, duration of OCD. The severity of OCD was assessed we will not report on them here. Novelty-seeking is using a clinician-administered version of the Yale– hypothesized to be a heritable tendency towards intense 162 P. Alonso et al. / Psychiatry Research 157 (2008) 159–168

Table 1 The validated Spanish version of the TCI was Sociodemographic and clinical characteristics of 60 patients with employed in the present study (Gutierrez et al., 2001). obsessive–compulsive disorder and healthy comparison subjects It showed good psychometric properties including its Characteristics Subjects with Healthy comparison α a internal consistency (all Cronbach's coefficients were OCD (n=60) subjects (n=60) above 0.65, except Persistence, α=0.49), concurrent Age, years 31.8±9.5 31.6±9.1 (18–55) validity and factorial structure. [mean±S.D. (range)] (18–55) Male/female 37/23 37/23 Education, years 10.9±2.7 11.0±2.8 Table 2 [mean±S.D. (range)] Sociodemographic and clinical characteristics of 60 patients with – Age at onset of OCD, years 17.9±5.8 obsessive compulsive disorder [mean±S.D. (range)] (7–40) Characteristics Subjects with OCD P Duration of illness, years 13.7±8.9 [mean±S.D. (range)] (1–39) Treated for a Newly minimum period recruited Global Y-BOCS score 19.7±12.2 (mean±S.D.) of 1 year (n=38) (n=22) Y-BOCS obsessions score 10.1±6.0 Age, years 33.0±8.4 29.8±11.0 0.2 (mean±S.D.) [mean±S.D. (range)] (20–52) (18–55) Y-BOCS compulsions score 9.6±6.3 Male/female 24/14 13/9 0.7 (mean±S.D.) Age at onset of OCD, 17.6±4.4 18.4±7.8 0.6 HDRS score (mean±S.D.) 12.1±5.5 years (mean±S.D.) OCD=obsessive–compulsive disorder, Y-BOCS=Yale–Brown Ob- Duration of illness, years 15.0±8.3 11.4±9.7 0.1 (mean±S.D.) sessive–Compulsive Scale, HDRS=Hamilton Depression Rating Scale. Pre-treatment Y-BOCS a There were no significant differences in age, gender or education score (mean±S.D.) between subjects with OCD and healthy comparison subjects. Global 26.5±7.9 Obsessions 13.6±3.4 Compulsions 12.9±5.2 exhilaration or excitement in response to novel stimuli or Pre-treatment HDRS 12.4±5.8 cues for potential rewards or potential relief of punish- score (mean±S.D.) ment. It involves impulsiveness in decision-making, Treatment status: SRI trials completed, extravagance in approach to cues of reward and disorder- N (%) liness with quick loss of temper. Harm avoidance reflects 1 12 (31.6%) a tendency to respond intensely to signals of aversive 2 13 (34.2%) stimuli, and thereby the subject learns to inhibit behaviour 3 11 (28.9%) to avoid punishment, novelty and frustrative nonreward. It >3 2 (5.3%) Complete CBT 28 (73.7%) involves anticipatory worry about possible problems, fear protocol, N (%) of uncertainty, shyness with strangers, and easy fatigabil- Reduction in Y-BOCS 47.0±33.2 ity. Reward dependence reveals a tendency to respond global score after intensely to signals of reward (particularly verbal signals treatment (mean %±S.D.) of social approval, sentiment and succour) and to maintain Y-BOCS score at TCI assessment (mean±S.D.) or resist extinction of behaviour that has previously been Global 13.5±10.6 30.5±5.5 <0.001 associated with rewards or relief from punishment. It is Obsessions 7.1±5.2 15.3±2.9 <0.001 manifest as sentimentality, social attachment and depen- Compulsions 6.3±5.5 15.2±2.6 <0.001 dence upon others' approval. Persistence reflects a HDRS at TCI assessment 10.4±5.4 15.1±3.7 0.001 tendency to industriousness, ambitious overachieving (mean±S.D.) Personality dimensions: and perseverance despite . Self-directedness Harm avoidance 67.9±9.8 71.5±6.4 0.09 refers to self-determination and “willpower”, or the ability Novelty-seeking 43.2±8.2 43.9±11.3 0.7 of an individual to control, regulate and adapt behaviour Reward dependence 48.1±8.9 51.8±8.3 0.1 to fit the situation in accord with individually chosen goals Persistence 41.7±8.7 42.0±9.8 0.8 and values. Cooperativeness can be formulated as a Self-directedness 38.2±11.6 34.5±11.4 0.2 Cooperativeness 41.9±9.1 45.0±9.8 0.2 developmental process which includes social acceptance, Self-transcendence 36.3±9.2 37.1±13.1 0.7 empathy, helpfulness, compassion and pure-hearted OCD=obsessive–compulsive disorder; Y-BOCS=Yale–Brown Ob- principles. Self-transcendence reflects a tendency toward sessive–Compulsive Scale, HDRS=Hamilton Depression Rating spirituality and identification with the wider world, as Scale, SRI=serotonin reuptake inhibitors, CBT=cognitive–beha- well as the ability to accept ambiguity and uncertainty. vioural therapy. P. Alonso et al. / Psychiatry Research 157 (2008) 159–168 163

Table 3 Multiple linear regression analyses (stepwise method) – Frequencies of the major symptom dimensions of the Yale Brown were conducted to assess whether temperament and Obsessive–Compulsive Checklist in 60 patients with obsessive– compulsive disorder character patterns were related to the presence of specific obsessive–compulsive symptom dimensions. In these Absent Present Major a models, the patients' scores on each of the subscales of symptom symptom symptom the TCI were entered as independent variables and the n % n % n % scores on the five previously identified OCD symptom Hoarding 45 75.0 11 18.3 4 6.7 dimensions as dependent variables. To control for the Aggressive/Checking 16 26.7 18 30.0 26 43.3 effect of symptom severity and depression, all analyses Contamination/Cleaning 29 48.3 14 23.3 17 28.3 Sexual/Religious 44 73.3 7 11.7 9 15.0 were repeated by entering the total Y-BOCS and HDRS Symmetry/Ordering 41 68.3 12 20.0 7 11.7 scores first in the models (enter method). a More than one principal category was allowed for each patient. The SPSS statistical package (version 10.0) was used in all analyses. The significance level was set at P <0.05 (2-tailed) except for those analyses where the A computer program was used that provided raw use of Bonferroni correction raised the significance scores and calculated T scores according to norms for the level to 0.007. seven scales (Cloninger et al., 1994). Raw scores were calculated by adding 1 point for each item answered in 3. Results the direction predicted by the scale. Standardization to T scores adjusts mean scores of each scale to the Demographic and clinical characteristics of the patient reference value of 50 points. and control groups are shown in Tables 1 and 2.Data comparability analyses showed no differences in sex 2.4. Statistical analysis distribution, age, age at onset of OCD, previous duration of the disorder, and personality dimensions assessed by The TCI scores on temperament and character the TCI between patients who had completed a minimum dimensions were compared between OCD patients and treatment period of 1 year and those newly referred for healthy subjects using an independent t test. To control therapy. Pharmacological and behavioural treatment for error derived from multiple comparisons, the received by the group of 38 patients in the long-term Bonferroni correction was employed (significance follow-up study is summarized in Table 2. At the time of level was established at 0.007 when comparing both personality assessment, subjects who had undergone groups on the seven personality dimensions). Relation- standardized OCD treatment in our unit scored signifi- ships between personality dimensions and clinical cantly lower than the newly recruited patients on the Y- variables such as age, age at onset of OCD, previous BOCS global, obsessions and compulsions subscales as duration of the disorder, severity of OCD and comorbid well as on the HDRS (Table 2). Frequencies of the depression were initially explored with Pearson corre- principal OCD symptom dimensions in the sample as a lation coefficients. Due to the detected influence of age whole were distributed as shown in Table 3. and level of depression on TCI scores, partial correlation Seventeen patients (28.3 %) fulfilled criteria for other coefficients between personality variables and clinical comorbid psychiatric disorders, including Major De- factors were then calculated controlling for age and pressive Disorder (N=10), Dysthymic Disorder (N=3), HDRS scores. Anxiety Disorders other than OCD (N=1) and Eating

Table 4 Comparison of Temperament and Character Inventory (TCI) scores between obsessive–compulsive patients and healthy comparison subjects Personality dimensions OCD, n=60 (mean±S.D.) Controls, n=60 (mean±S.D.) tP 95% CI Harm avoidance 69.2±8.8 51.9±8.7 10.7 <0.001 14.1 20.5 Novelty seeking 43.4±9.4 52.0±10.7 −4.6 <0.001 −12.2 −4.9 Reward dependence 49.5±8.8 51.0±8.4 −0.9 0.32 −4.6 1.5 Persistence 41.8±9.1 44.5±10.2 −1.5 0.1 −6.1 0.82 Self-directedness 36.8±11.6 52.6±8.4 −8.5 <0.001 −19.4 −12.1 Cooperativeness 43.0±9.4 50.6±6.7 −5.0 <0.001 −10.5 −4.6 Self-transcendence 36.6±10.7 39.0±10.9 −1.1 0.2 −6.2 1.5 OCD=obsessive–compulsive disorder. 164 P. Alonso et al. / Psychiatry Research 157 (2008) 159–168

Disorders (N=3). In all cases, OCD onset was prior to β=0.27, t=2.14, P =0.03). These results remained the development of any other psychiatric condition. unchanged when total Y-BOCS and HDRS scores Seven patients (11.6%) exhibited chronic motor tic were forced first into the models. None of the other disorder or Tourette's syndrome. OCD symptom dimensions were significantly related to Mean scores and standard deviations (S.D.) on TCI specific personality dimensions. subscales in OCD patients and healthy controls are presented in Table 4. Independent t tests indicated that 4. Discussion OCD subjects had significantly higher mean scores in harm avoidance (t=10.7, df=118, P<0.001) and lower The aim of this study was to examine the personality mean scores in novelty-seeking (t= −4.6, df=118, profile of OCD patients and to determine whether P <0.001), self-directedness (t = − 8.5, df=118, temperament and character dimensions were related to P <0.001) and cooperativeness (t = −5.0, df=118, different clinical aspects such as severity of obsessive– P<0.001) compared with healthy comparison subjects. compulsive or depressive symptoms, age at onset of the Since approximately one-third of OCD subjects had disorder, presence/absence of comorbid conditions, and comorbid psychiatric conditions, mainly affective dis- OCD symptom subtypes. orders, which have been reported to be related to spe- Our results are consistent with previous reports on cific TCI patterns (Joffe et al., 1993) and could therefore patterns of temperament and character in OCD subjects influence our results, we decided to repeat analyses on characterized by high scores on the harm avoidance TCI dimensions just considering differences between dimension (Pfohl et al., 1990; Richter et al., 1996; Bejerot OCD patients without comorbid psychiatric disorders et al., 1998; Kusunoki et al., 2000; Lyoo et al., 2001)and and healthy controls. Results of this second analysis low scores on the novelty-seeking (Kusunoki et al., 2000; were identical to those obtained in the sample as a Lyoo et al., 2001), self-directedness (Bejerot et al., 1998; whole, indicating that pure OCD subjects had signifi- Kusunoki et al., 2000; Lyoo et al., 2001) and coopera- cantly higher mean scores in harm avoidance (68.2 ±9.4 tiveness (Bejerot et al., 1998; Kusunoki et al., 2000) vs. 51.9±8.7; t=8.8, df=99, P<0.001) and lower mean dimensions. Recent cognitive theories emphasize the scores in novelty-seeking (43.3±8.0 vs. 52.0±10.7; t= important role of certain belief domains in the develop- −4.3, df=99, P<0.001), self-directedness (38.6±12.1 ment of OCD including an exaggerated sense of vs. 52.6±8.4; t=−6.4, df=99, P<0.001) and coopera- responsibility and the overestimation of threat (Steketee tiveness (42.3 ±10.4 vs. 50.6±6.7; t=−4.5, df=99, et al., 1998). Distorted appraisals of the power of one's P<0.001) than healthy comparison subjects. actions to produce or prevent harm combined with an No gender differences on TCI scores were detected in unusually high estimation of risk would explain why most the OCD group, except for the self-transcendence OCD sufferers seem to view situations as dangerous subscale, in which women scored significantly higher unless proven safe and become highly vigilant when than men (40.6±11.6 vs. 34.1±9.4, t=−2.3, df=58, novel circumstances arise. Excessive anticipatory worries P=0.02). about possible problems or the tendency to inhibit A significant negative correlation was detected bet- exploratory activity that could lead to novel potentially ween age and NS (r=−0.52, P<0.001) in the OCD dangerous or punishable situations, both characteristics of group. No significant correlations between TCI scores subjects with high harm avoidance and low novelty- and age at onset of OCD or previous duration of the seeking scores, can be considered congruent with this disorder were detected when controlling for current age. hypothesized cognitive schema. Presence of depressive symptoms assessed by the Results on character dimensions have been interpreted HDRS was closely correlated with HA (r =0.49, from different perspectives. Low self-directedness has P<0.001). These results remained unchanged when been postulated as being related to a reduced ability of controlling for OCD severity and age. OCD severity OCD patients to regulate their own thoughts and be- assessed by the Y-BOCS appeared to be correlated with haviors, especially when they seek to guide them towards HA (r=0.39, P=0.002) and NS scores (r=−0.28, concrete goals, due to the presence of invasive obsessions P=0.002), but these correlations were not statistically and compulsions (Lyoo et al., 2001). On the other hand, significant when controlling for age and presence of low self-directedness and cooperativeness have been depressive symptoms. hypothesized to reflect the high frequency of comorbid In the OCD group, multiple linear regression personality disorders observed in OCD patients, since low analyses revealed weak partial correlations between scores on both dimensions are common characteristics of scores on the hoarding dimension and HA (R2 =0.07, Axis II pathology (Bejerot et al., 1998). Finally, Cloninger P. Alonso et al. / Psychiatry Research 157 (2008) 159–168 165 postulates that character can be understood in terms of the tients with these personality profiles: excessive antici- epigenetic development of increasingly inclusive concepts patory worry about possible problems and fear of of the self: identification as an autonomous individual uncertainty are characteristic of subjects with obses- (self-directedness), as an integral part of human society sive–compulsive personality, while shyness with stran- (cooperativeness) and as an integral part of the universe gers and intense response to signals of aversive stimuli (self-transcendence) (Cloninger et al., 1993). The presence can be considered the core symptoms of avoidant per- of pervasive intrusive thoughts since childhood or sonality disorder. Our results would further support the along with accompanying feelings of isola- hypothesis that differential relationships exist among tion, guilt and blame, typical of OCD sufferers, might personality dimensions and specific OCD symptom render such adequate character development difficult. profiles (Baer, 1994; Mataix-Cols et al., 2000). So, Although initial exploratory analysis appeared to previous conflicting findings on the role of personality indicate that OCD severity was correlated with scores factors in OCD may be due in part to differences in the on HA and NS dimensions, these results were not sta- constitution of the patients groups studied, i.e. the tistically significant after controlling for age and presence differential presence of patients with hoarding symptoms. of depressive symptoms. On the other hand, a strong Our results must be interpreted with several caveats in correlation between depressive mood symptoms and HA mind. A personality profile characterized by high HA scores, regardless of OCD severity, was detected in our scores has been described to be related to different anxiety sample. In the only previous study addressing this issue, and affective disorders (Ball et al., 2002). So, the use of a Lyoo et al. (2001) reported that high HA and low SD psychiatric control group would have been interesting to scores significantly predicted more severe obsessive– elucidate whether the detected relation between person- compulsive symptoms, controlling for age, gender and ality dimensions and OCD is a specific one. level of depression. Nevertheless none of these patients Although the dimensional approach adopted in this suffered from a Major Depressive Episode, since the study has the potential advantage of overcoming the presence of comorbid Axis I disorders was considered an difficulty of recruiting a sufficient sample size for each exclusion criterion by the authors. Thirteen of our patients OCD clinical subtype, some symptom dimensions (i.e. (21.6%) fulfilled criteria for Major Depressive or sexual/religious, symmetry/ordering, hoarding) were Dysthymic Disorder, a percentage similar to usually present in a reduced proportion of patients. So, the reported rates of comorbid depression in OCD (Black and sample size might have been insufficient to detect a Noyes, 1990). Absence of comorbid mood disorders in significant relationship between some of these clinical Lyoo et al.'s sample may, at least partially, account for dimensions and temperament and character profile. The differences between studies on the relationship between results of the current study need to be replicated in larger personality and OCD severity. Our findings are con- samples to address this issue as well as to confirm the sistent with the previously observed influence of mood stability of the detected association between hoarding states on personality assessment, especially regarding the and HA scores. harm avoidance dimension (Svrakic et al., 1992; On the other hand, differentiating hoarding as an Strakowski et al., 1992; Joffe et al., 1993). In fact, in a OCD symptom from hoarding as a personality trait study designed to assess changes on TCI scores in associated to obsessive–compulsive personality disor- response to OCD therapy, Lyoo et al. (2003) reported that der (OCPD) is not always easy, and this fact could have the decrease in levels of depressive symptoms was the influenced our results on the proposed relationship major contributing factor for the change in HA scores between hoarding and some personality dimensions, after treatment. All these findings point to the importance specially high HA scores. Since no psychometrically of considering the presence of comorbid depressive sound hoarding scales were available at the time of this symptoms in assessing personality dimensions in OCD study, we tried to overcome this limitation employing patients. the strict definition of hoarding proposed by Frost and Our results on the relationship between high HA and Hartl (1996), which appears to show no correlation with the development of hoarding symptoms are consistent a global measure of obsessive–compulsive personality with previous reports on the association between disorder as well as with the other characteristics of this hoarding and comorbid personality disorders. In the except for perfectionism (Frost and study by Mataix-Cols et al. (2000), obsessive–compul- Gross, 1993). Reliable and valid instruments recently sive and avoidant were the personality disorders most developed to assess hoarding symptoms (Frost et al., closely related to the hoarding dimension. Several 2004) should be used in future studies to clarify the aspects of the HA dimension are common among pa- relationship between hoarding and personality. 166 P. Alonso et al. / Psychiatry Research 157 (2008) 159–168

Studies on comorbidity, symptom pattern and etiolog- controls. No significant differences were detected ical factors suggest that OCD is far from being a between the groups in scores for novelty-seeking, reward homogeneous disorder. Although scarce data are avail- dependence, persistence or character factors. Similar to able on this issue in adult samples, temperamental Cloninger's HA concept, Kagan (1997) defined a heterogeneity in child and adolescent OCD has been temperament construct, named “behavioural inhibition observed by Ivarsson and Winge-Westholm (2004),who to the unfamiliar (BI)”, which reflects the tendency to be described two temperamental subgroups using hierarchi- unusually shy and fearful as a toddler and quiet and cal cluster analysis in a sample of 83 obsessive children withdrawn in unfamiliar situations in the preschool and and adolescents: the uninhibited and the inhibited/shy early school-age years. Several studies have described groups. The authors also hypothesize the existence of an that children classified as “stable inhibited”–those who “autistic-like” sub-group included in the inhibited/shy remained inhibited throughout childhood – have an one, characterized by low levels of sociability and increased risk for anxiety disorders (Hirshfeld et al., presence of repeating compulsions. Such a study of 1992). Specific anxiety disorders related to BI appear to temperamental heterogeneity in our sample of adults with change from early childhood to later years, since OCD is currently underway. younger children had mainly avoidant, overanxious Our results support the importance of considering the and phobic disorder, while 3 years later they had more presence of comorbid depressive symptoms in assessing avoidant and separation anxiety disorder (Biederman et personality dimensions in OCD. Notwithstanding, some al., 1993). Due to the low prevalence of OCD in pediatric OCD patients suffer from other comorbid conditions years, no significant association between BI and OCD besides depression and anxiety disorders, such as tic could be detected in these studies. Nevertheless, Ivarsson disorder, Tourette's syndrome or autistic traits similar to and Winge-Westholm (2004), specifically assessed the those present in Asperger's disorder, that appear to be relation between OCD and temperament in a group of 83 related to specific symptom dimensions or genetic obsessive children and adolescents and detected higher findings (Mataix-Cols et al., 2005). The relation of scores on the “shyness” subscale and lower scores on the personality dimensions with these comorbid conditions “activity” subscale of the EAS (Emotionality, Activity, could not be assessed in our study due to the absence of a Sociability) questionnaire in the OCD group compared sufficient number of afflicted patients. Future studies on with normal controls. The authors stated that, together, the relation between personality dimensions and specific these scores could be seen as indicative of the presence of comorbid conditions in OCD might help to clarify the behavioural inhibition. Adults characterized by high remarkable heterogeneity of the disorder. harm avoidance and low novelty-seeking scores on Finally, any discussion of the relationship between Cloninger's TCI exhibit excessive anticipatory worries personality and OCD must be extremely cautious in about possible problems, fear of uncertainty, shyness hypothesizing on the direction of causality between both with strangers and a tendency to inhibit exploratory phenomena. Personality patterns may render an individ- activity that could lead to novel potentially dangerous or ual vulnerable to the development of obsessive– punishable situations. These behavioural characteristics compulsive disorder, but the presence of debilitating are quite similar to those ascribed to inhibited children and invasive obsessive–compulsive symptoms from late who are described as quiet, fearful, withdrawn and adolescence or early adulthood may irrevocably change noninteractive in unfamiliar situations. So, although not the individual's habitual patterns of thought, behaviour identical, harm avoidance, novelty-seeking and beha- and emotional regulation. Moreover, in the case of such a vioural inhibition can be seen as related theoretical lifelong condition as OCD, the disorder itself becomes a concepts, addressing an innate temperamental tendency stable part of the experience and behavioural repertoire that describes one's style of behaviour in situations of of the individual, and discerning which symptoms are uncertainty. Studies comparing the EAS and the those of OCD and which more accurately reflect Cloninger systems would help to clarify the exact personality may be quite difficult. Studies in children relation between these theoretical models. On the other and adolescents with OCD, who have been suffering hand, prospective studies following inhibited children from the disorder for less time, might help to clarify this into adulthood would help to resolve whether certain issue. Gothelf et al. (2004), in the only published study temperament factors really reflect an enhanced vulner- investigating Cloninger's temperament and character ability to develop OCD or other anxiety disorders. model in children with OCD and other anxiety disorders In summary, the results support the existence of a (AD), reported higher scores on the HA dimension in dimensional personality profile associated with OCD children with OCD and other AD compared with normal and characterized by high harm avoidance and low P. Alonso et al. / Psychiatry Research 157 (2008) 159–168 167 novelty-seeking, self-directedness and cooperativeness Cloninger, C.R., Svrakic, D.M., Przybeck, T.R., 1993. A psychobi- scores. This pattern of temperament and character was ological model of temperament and character. Archives of General Psychiatry 50, 975–990. not related to OCD severity, age at onset or previous Cloninger, C.R., Przybeck, T.R., Svrakic, D.M., Wetzel, R.D., 1994. duration of the disorder, but partially to the presence of The Temperament and Character Inventory (TCI): A Guide to its comorbid depressive symptoms. After all state variables Development and Use. Center for Psychobiology of Personality, were controlled, hoarding obsessions and compulsions St. Louis. were significantly associated with high scores on the First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.BW., 1997. Structured Clinical Interview for DSM-IV Axis I Disorders– harm avoidance dimension. Probably, much of the lack Clinician Version (SCID–CV). American Psychiatric Press, of consensus in the literature on personality factors Washington, DC. associated with OCD, both categorically and dimen- Freud, S., 1908. Character and anal erotism. In Strachey, J. (Ed. and sionally considered, may have arisen from a failure to Trans.), Standard Edition of the Complete Psychological Works of consider the clinical heterogeneity of the disorder. Future Sigmund Freud. Hogarth Press, London. Frost, R.O., Gross, R.C., 1993. The hoarding of possessions. Behaviour research should take into account the influence of Research and Therapy 31, 367–381. comorbid clinical conditions or symptom subtypes in Frost, R.O., Hartl, T., 1996. A cognitive–behavioral model of addressing the role of personality factors in obsessive– compulsive hoarding. Behaviour Research and Therapy 34, compulsive disorder. 341–350. Frost, R.O., Steketee, G., Grisham, J., 2004. Measurement of compulsive hoarding: saving inventory-revised. Behaviour Re- Acknowledgements search and Therapy 42, 1163–1182. Fullana, M.A., Mataix-Cols, D., Caseras, X., Alonso, P., Menchon, J.M., This research was supported in part by grant FIS 01/ Vallejo, J., Torrubia, R., 2004. High sensitivity to punishment and 1558 and 02/0102 from the Spanish Ministerio de low impulsivity in obsessive–compulsive patients with hoarding – Sanidad y Consumo and grant 010210 from Fundació symptoms. Psychiatry Research 129, 21 27. Goodman, W.K., Price, L.H., Rasmussen, S.A., Mazure, C., La Marató TV3. PA was funded by Fundació Pi i Fleischmann, R.L., Hill, C.L., Heninger, G.R., Charney, D.S., Sunyer, Hospital Universitario de Bellvitge. 1989. The Yale–Brown Obsessive–Compulsive Scale (Y-BOCS): Part I. Development, use and reliability. Archives of General Psychiatry 46, 1006–1011. References Gothelf, D., Aharonovsky, O., Horesh, N., Carty, T., Apter, A., 2004. Life events and personality factors in children and adolescents with Alonso, P., Menchon, J.M., Pifarre, J., Mataix-Cols, D., Torres, L., obsessive–compulsive disorder and anxiety disorders. Compre- Salgado, P., Vallejo, J., 2001. Long-term follow-up and predictors hensive Psychiatry 45, 192–198. of clinical outcome in obsessive–compulsive patients treated with Gutierrez, F., Torrens, M., Boget, T., Martin-Santos, R., Sangorrin, J., serotonin reuptake inhibitors and behavioural therapy. Journal of Perez, G., Salamero, M., 2001. Psychometric properties of the Clinical Psychiatry 62, 535–540. Temperament and Character Inventory (TCI) questionnaire in a American Psychiatric Association, 1994. Diagnostic and Statistical Spanish psychiatric population. Acta Psychiatrica Scandinavica Manual of Mental Disorders, 4th Edition. American Psychiatric 103, 143–147. Association Press, Washington, DC. Hamilton, M., 1960. A rating scale for depression. Journal of Baer, L., 1994. Factor analysis of symptom subtypes of obsessive– Neurology, Neurosurgery and Psychiatry 23, 56–61. compulsive disorder and their relation to personality and tic Hirshfeld, D.R., Rosenbaum, J.F., Biederman, J., Bolduc, E.A., disorders. Journal of Clinical Psychiatry 55 (Suppl. 3), 18–23. Faraone, S.V., Snidman, N., Reznick, J.S., Kagan, J., 1992. Stable Ball, S., Smolin, J., Shekhar, A., 2002. A psychobiological approach to behavioral inhibition and its association with anxiety disorder. personality: examination within anxious outpatients. Journal of Journal of the American Academy of Child and Adolescent Psychiatric Research 36, 97–103. Psychiatry 31, 103–111. Bejerot, S., Schlette, P., Ekselius, L., Adolfsson, R., von Knorring, L., Horesh, N., Dolberg, O.T., Kirschbaum-Aviner, N., Kotler, M., 1997. 1998. Personality disorders and relationship to personality dimen- Personality differences between obsessive–compulsive disorder sions measured by the Temperament and Character Inventory in subtypes: washers versus checkers. Psychiatry Research 71, 197–200. patients with obsessive–compulsive disorder. Acta Psychiatrica Ivarsson, T., Winge-Westholm, C., 2004. Temperamental factors in Scandinavica 98, 243–249. children and adolescents with obsessive–compulsive disorder Biederman, J., Rosenbaum, J.F., Bolduc-Murphy, E.A., Faraone, S.V., (OCD) and in normal controls. European Child and Adolescent Chaloff, J., Hirshfeld, D.R., Kagan, J., 1993. A 3-year follow-up of Psychiatry 13, 365–372. children with and without behavioral inhibition. Journal of the Janet, P., 1903. Les Obsessions et la Psychasténie. Alcan, Paris. American Academy of Child and Adolescent Psychiatry 32, 814–821. Joffe, R.T., Bagby, M., Levitt, A.J., Regan, J.J., Parker, J.D.A., 1993. Black, D., Noyes, R., 1990. Comorbidity and obsessive–compulsive The Tridimensional Personality Questionnaire in major depression. disorder. In: Maser, J., Cloninger, C. (Eds.), Comorbidity of Mood and American Journal of Psychiatry 150, 959–960. Anxiety Disorders. American Psychiatric Press, Washington, DC. Kagan, J., 1997. Temperament and the reactions to unfamiliarity. Child Cloninger, C.R., 1987. A systematic method for clinical description Development 68, 139–144. and classification of personality variants. A proposal. Archives of Kusunoki, K., Sato, T., Taga, C., Yoshida, T., Komori, K., Narita, T., General Psychiatry 44, 573–588. Hirato, S., Iwata, N., Ozaki, N., 2000. Low novelty-seeking 168 P. Alonso et al. / Psychiatry Research 157 (2008) 159–168

differentiates obsessive–compulsive disorder from major depres- Pfohl, B., Black, D., Noyes, R., Kelley, M., Blum, N., 1990. A test of sion. Acta Psychiatrica Scandinavica 101, 403–405. the tridimensional personality theory: association with diagnosis Leckman, J.F., Grice, D.E., Boardman, J., Zhang, H., Vitale, A., and platelet imipramine binding in obsessive–compulsive disorder. Bondi, C., Alsobrook, J., Peterson, B.S., Cohen, D.J., Rasmussen, Biological Psychiatry 28, 41–46. S.A., Goodman, W.K., McDougle, C.J., Pauls, D.L., 1997. Richter, M.A., Summerfeldt, L.J., Joffe, R.T., Swinson, R., 1996. The Symptoms of obsessive compulsive disorder. American Journal Tridimensional Personality Questionnaire in obsessive–compul- of Psychiatry 154, 911–917. sive disorder. Psychiatry Research 65, 185–188. Lyoo, I.K., Lee, D.W., Kim, Y.S., Kong, S.W., Kwon, J.S., 2001. Shtasel, D.L., Gur, R.E., Mozley, P.D., Richards, J., Taleff, M.M., Patterns of temperament and character in subjects with obsessive– Heimberg, C., Gallacher, F., Gur, R.C., 1991. Volunteers for compulsive disorder. Journal of Clinical Psychiatry 62, 637–641. biomedical research. Recruitment and screening of normal Lyoo, I.K., Yoon, T., Kang, D.H., Kwon, J.S., 2003. Patterns of changes controls. Archives of General Psychiatry 48, 1022–1025. in Temperament and Character Inventory scales in subjects with Spitzer, R.L., Williams, J.B.W., Gibbon, M., 1989. Structured Clinical obsessive–compulsive disorder following a 4-month treatment. Interview for DSM-III-R: Non-Patient Version (SCID-NP). New Acta Psychiatrica Scandinavica 107, 298–304. York State Psychiatric Institute, New York, NY. Mataix-Cols, D., Rauch, S.L., Manzo, P.A., Jenike, M.A., Baer, L., Steketee, G.S., Frost, R.O., Rheaume, J., Wilhelm, S., 1998. Cognitive 1999. Use of factor-analyzed symptom dimensions to predict theory and treatment of obsessive–compulsive disorder. In: Jenike, outcome with serotonin reuptake inhibitors and placebo in the M.A., et al. (Eds.), Obsessive–Compulsive Disorders. Practical treatment of obsessive–compulsive disorder. American Journal of Management, 3rd ed. Mosby, Inc., St. Louis, pp. 368–399. Psychiatry 156, 1409–1416. Strakowski, S.M., Faedda, G.L., Tohen, M., Goodwin, D.C., Stoll, A.L., Mataix-Cols, D., Baer, L., Rauch, S.L., Jenike, M.A., 2000. Relation 1992. Possible affective-state dependence of the Tridimensional of factor-analyzed symptom dimensions of obsessive–compulsive Personality Questionnaire in first-episode psychosis. Psychiatry disorder to personality disorders. Acta Psychiatrica Scandinavica Research 41, 215–227. 102, 1–4. Summerfeldt, L.J., Huta, V., Swinson, R.P., 1998. Personality and Mataix-Cols, D., Rauch, S.L., Baer, L., Eisen, J.L., Shera, D.M., obsessive–compulsive disorder. In: Swinson, R.P., et al. (Eds.), Goodman, W.K., Rasmussen, S.A., Jenike, M.A., 2002. Symptom Obsessive–Compulsive Disorder. Theory, Research and Treat- stability in adult obsessive compulsive disorder: data from a ment. The Guilford Press, New York, pp. 79–119. naturalistic two-year follow-up study. American Journal of Summerfeldt, L.J., Richter, M.A., Antony, M.M., Swinson, R.P., 1999. Psychiatry 159, 263–268. Symptom structure in obsessive–compulsive disorder: a confir- Mataix-Cols, D., do Rosario-Campos, M., Leckman, J.F., 2005. A matory factor-analytic study. Behaviour Research and Therapy 37, multidimensional model of obsessive–compulsive disorder. Amer- 297–311. ican Journal of Psychiatry 162, 228–238. Svrakic, D.M., Przybeck, T.R., Cloninger, R., 1992. Mood states and Matsunaga, H., Kiriike, N., Matsui, T., Iwasaki, Y., Koshimune, K., personality traits. Journal of Affective Disorders 24, 217–226. Ohya, K., Stein, D.J., 2001. A comparative study of clinical features between pure checkers and pure washers categorized using a lifetime symptom rating method. Psychiatry Research 105, 221–229.